Chromophore Distortions in Photointermediates of Proteorhodopsin

Oct 13, 2017 - Proteorhodopsin (PR) is the most abundant retinal protein on earth and functions as a light-driven proton pump. Despite extensive effor...
0 downloads 12 Views 5MB Size
Download PDF
Article Cite This: J. Am. Chem. Soc. 2017, 139, 16143-16153

pubs.acs.org/JACS

Chromophore Distortions in Photointermediates of Proteorhodopsin Visualized by Dynamic Nuclear Polarization-Enhanced Solid-State NMR Michaela Mehler,† Carl Elias Eckert,‡ Alexander J. Leeder,§ Jagdeep Kaur,† Tobias Fischer,‡ Nina Kubatova,† Lynda J. Brown,§ Richard C. D. Brown,§ Johanna Becker-Baldus,† Josef Wachtveitl,‡ and Clemens Glaubitz*,† †

Institute for Biophysical Chemistry & Centre for Biomolecular Magnetic Resonance, Goethe-University Frankfurt, Frankfurt 60438, Germany ‡ Institute for Physical and Theoretical Chemistry, Goethe-University Frankfurt, Frankfurt 60438, Germany § Department of Chemistry, University of Southampton, Southampton SO17 1BJ, United Kingdom S Supporting Information *

ABSTRACT: Proteorhodopsin (PR) is the most abundant retinal protein on earth and functions as a light-driven proton pump. Despite extensive efforts, structural data for PR photointermediate states have not been obtained. On the basis of dynamic nuclear polarization (DNP)-enhanced solid-state NMR, we were able to analyze the retinal polyene chain between positions C10 and C15 as well as the Schiff base nitrogen in the ground state in comparison to light-induced, cryotrapped K- and M-states. A high M-state population could be achieved by preventing reprotonation of the Schiff base through a mutation of the primary proton donor (E108Q). Our data reveal unexpected large and alternating 13 C chemical shift changes in the K-state propagating away from the Schiff base along the polyene chain. Furthermore, two different M-states have been observed reflecting the Schiff base reorientation after the deprotonation step. Our study provides novel insight into the photocycle of PR and also demonstrates the power of DNP-enhanced solid-state NMR to bridge the gap between functional and structural data and models.

■

INTRODUCTION Rhodopsins are found in all phyla of life and represent the most abundant phototrophic systems.1−4 By far the most abundant family members belong to the proteorhodopsins (PRs), which provided the first evidence for a bacterial retinal-based photoreceptor.1 Blue and green light-absorbing subfamilies have been found.5 Their identification through metagenomic screens of uncultured sea samples has led to many hundreds of PR-like sequences distributed in numerous microorganisms from different geographic areas.6−8 Their prevalent occurrence in microbial communities in the ocean’s photic zone and their ability to act as light-driven proton pumps, which creates a transmembrane electrochemical gradient, makes retinal-based phototrophy a very important bioenergetic factor in marine ecosystems during nutrient deficient periods.9,10 These studies also triggered the discovery of new rhodopsins and helped to expand knowledge of their spectrum of functions and distribution. PR features the typical heptahelical transmembrane bundle with a retinal chromophore covalently bound to Lys231 and is a DTE-motif ion pump with the primary proton acceptor Asp97 and proton donor Glu108.11−13 Furthermore, Asp227 and Arg94 form together with the proton acceptor the Schiff base’s © 2017 American Chemical Society

counterion complex. Glu142 is proposed to be involved in proton release.14 In addition, proton-pumping efficiency is pH regulated due to the high pKa of Asp97, which is close to the environmental pH. In contrast to other microbial rhodopsins, PR forms pentameric complexes supportive of its function.14−16 For proton transfer to be performed, PR undergoes a lightinduced photocycle with a number of characteristic intermediate states (Figure 1A).6,10 Different from archaeal bacteriorhodopsin, only the all-trans retinal configuration is present in the dark, and no dark-light adaptation step is observed. The photocycle is initiated by retinal isomerization from all-trans,15-anti to 13-cis,15-anti around the C13C14 bond, after which the K-state intermediate is formed. Subsequently, the protonated Schiff base (pSB) transfers its proton to Asp97, leading to the M-state with a reduced πelectron delocalization, which yields a characteristic blue shift of the protein’s absorption wavelength of 410 nm. Next, the SB is reprotonated by the proton donor Glu108 during the N-state. From there, the retinal reisomerizes back to all-trans in the Ostate. Finally, PR returns to the ground state with proton Received: May 16, 2017 Published: October 13, 2017 16143

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

Figure 1. (A) The PR photocycle is initiated by light-induced isomerization of retinal from all-trans,15-anti to 13-cis,15-anti, leading to the K-state (in the range of picoseconds). Next, the Schiff base transfers its proton to the primary proton acceptor Asp97, resulting in the blue-shifted M-state. Subsequently, the Schiff base is reprotonated from the proton donor Glu108, leading to the formation of the N-intermediate. In the last two steps, the retinal reisomerization takes place (O-state) followed by proton uptake from the cytoplasma and reprotonation of E108.10 The cycle is completed after approx. 250 ms (see Figure 4E for further details). (B) Two in situ illumination schemes used here for DNP-enhanced NMR measurements. Samples can be illuminated directly within the MAS probe under cryogenic conditions or at room temperature followed by rapid freezing and subsequent sample transfer into the magnet. (C) Retinal labeling schemes 14,15-13C2-E-retinal (13C2-retinal) and 10,11,12,13,14,15-13C6-E-retinal (13C6-retinal).

spectroscopic characterization of the photocycle on the one hand and 3D structure analysis by crystallographic approaches on the other. ssNMR reports on the structure and electronic environment of the chromophore, and such data could be eventually linked to optical properties via quantum chemical approaches. However, only very few ssNMR studies have been reported on photointermediates due to the substantial challenges associated with the limited sensitivity of such experiments. Available data are restricted to bacteriorhodopsin and sensory and visual rhodopsin.29,34−37 Fortunately, the advent of dynamic nuclear polarization (DNP) enables new perspectives: The sensitivity of ssNMR can be boosted by orders of magnitude by magnetization transfer from suitable polarizing agents containing stable radicals to the nuclei of interest.38,39 For membrane protein samples, up to 50−60-fold sensitivity enhancements have been reported. Such an approach is of emerging importance for the detection of subpopulations of cryotrapped photointermediate states as demonstrated for bacteriorhodopsin and channelrhodopsin-2.15,25 Here, we present DNP-enhanced MAS NMR data on 13Clabeled retinal incorporated into 15N-Lys-labeled green PR in the ground, K-, and M-states. In situ illumination protocols were established, and the M-state population was enriched by additionally introducing the primary proton donor mutation E108Q, which does not cause any structural effects but elongates the photocycle by preventing the reprotonation of the Schiff base. For following changes within the chromophore during the photocycle, the HC14-C15H single bond dihedral angle in 14,15-13C2-E-retinal (13C2-retinal, Figure 1C) and 13C chemical shifts in 10,11,12,13,14,15-13C6-E-retinal (13C6-retinal, Figure 1C) were determined. Our data reveal the existence of two M-states. Large chemical shift and dihedral angle changes were observed in the trapped photointermediates and reflect structural changes within chromophore and binding pocket. The response of the retinal polyene chain toward illumination during the first half of the PR photocycle is discussed in the

uptake by Glu108 (Figure 1A; for a review, see Bamann et al.17). Although PR has been extensively studied by advanced biophysical methods including time-resolved optical and vibrational spectroscopy, via liquid- and solid-state NMR, and X-ray crystallography,10,11,14,15,17−21 very little is known so far regarding conformational changes in the chromophore itself and the surrounding residues in the binding pocket during the photocycle. No structural data on photointermediates have been reported. Kinetic data for the photocycle are usually interpreted by analogy comparison with the less abundant, more exotic but extensively studied archaeal homologue bacteriorhodopsin (BR) from Halobacterium salinarium.10,22 However, the photocycle of the two proteins show clear differences, e.g., number and lifetime of the intermediate states and the retinal configuration in the dark state. Further differences include a highly conserved histidine residue associated with the active site of PR,14,23,24 a different proton release group, and the lack of highly ordered water 402 between Schiff base and proton acceptor, which facilitates proton transfer in BR.14 In addition, the discovery of many new rhodopsins has shown that the typical structural scaffold provides the basis for a great variety of functions, including cation and anion pumps, sensors, and channels. These also manifest themselves in different optical properties, photocycle kinetics, or populated intermediate states, which indicate important differences in their energy landscapes. These fine details need to be understood to comprehend how apparently similar molecular designs can yield such a variety of functions. Solid-state NMR (ssNMR), especially in the form of magic angle sample spinning (MAS NMR), is a powerful method for the investigation of membrane proteins and has been extensively applied to rhodopsins.15,18,25−34 The possibility of performing experiments on light-induced and cryotrapped photointermediate states directly within the lipid bilayers provides unique opportunities to bridge the gap between the 16144

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

were illuminated directly within the MAS rotor under cryogenic conditions or at room temperature followed by fast freezing (thermal trapping) (Figure 1B). In all cases, sample transparency was identified to be the most crucial parameter for achieving a high trapping efficiency as it determines the size of the light-activated protein population in the lipid bilayer entering the photocycle.31 The best light penetration was achieved by evenly spreading a thin film of proteorhodopsin liposomes at a high lipid:protein ratio of 2 (w/w) on the inner surface of a transparent sapphire MAS rotor. Such a setup results in a greatly reduced total protein sample (1−1.5 mg) compared to that of a fully filled MAS rotor (12−20 mg), which makes the use of DNP-enhanced solid-state NMR indispensable. First, an efficient protocol for trapping the K-state was established by illuminating the sample directly within the DNP probe at approximately 100 K with a bright LED light source for 40 min (Figure 1B). As a result, the DNP-enhanced DQF spectra of 13C2 or 13C6-retinal show chemical shift changes as well as additional peaks (Figure 3A and B). For example, in the ground state, a chemical shift for C14 of 120.2 ppm is observed, whereas under the described illumination, a population at 117.7 ppm arises (Figure 3A, B). The 15N chemical shift of the pSB is not affected (Figure 5A). Assigning this thermally trapped state as the K-state is further confirmed by cryo-UV−vis spectra recorded under similar conditions (Figure 3C). The dark-light difference spectrum shows a decrease of the ground state absorption at 520 nm, and a population at 580 nm arises, which is characteristic for the K-intermediate. As estimated from the NMR peak intensities, approximately 30% of the sample could be trapped in the K-state. This incomplete conversion is probably caused by equilibrium between K- and ground states, by the limited quantum yield, and by a fraction of the sample that was not illuminated due to nonideal coupling of the light source into the DNP probe. Cryotrapping of the M-Intermediate. Different possibilities exist, in principle, for trapping the M-state. One option would be thermal relaxation from the K-state; another would be to populate the M-state at higher temperature followed by thermal trapping.29 A suitable NMR spectroscopic readout is provided by the 15N resonance of the Schiff base, which shifts by approximately 150 ppm upon deprotonation in the Mstate.29 Unfortunately, all attempts to convert wild-type PR from its ground state into the M-state have failed (Figure S1). This is most likely due to the low population of the Mintermediate in the PR photocycle, which has been reported previously11 and is also shown here by time-resolved absorption measurements (Figure 4A). All expected photointermediates

context of time-resolved optical spectroscopy and in light of data from other microbial rhodopsins.

■

RESULTS Cryotrapping of the K-Intermediate. Proteorhodopsin was 15N-Lys labeled and reconstituted with 13C2-retinal or 13C6retinal (Figure 1C) so that retinal and Schiff base could be monitored in differently trapped states. All experiments were performed on proteoliposomes. DNP-enhanced, double quantum-filtered (DQF) 13C spectra of 13C2- and 13C6-retinal in the PR ground state are shown in Figure 2. Using

Figure 2. (A) DNP-enhanced, double quantum-filtered (DQF) 13C spectra of PR with 13C2 and 13C6 retinal. DNP enhancement is approximately 50-fold. (B) DNP-enhanced DQ-SQ spectrum of 13C6retinal in the PR ground state. The assigned chemical shifts are summarized in Table 1.

AMUPOL40 as polarizing agent, a 50-fold signal enhancement for the 13C retinal resonances could be routinely achieved (Figure 2A). Chemical shifts were assigned based on double quantum−single quantum (DQ-SQ) correlation experiments41 (Figure 2B, Table 1), which were also verified by 13C−13C proton-driven spin diffusion (PDSD) correlation spectra (Figure 6B). For photocycle intermediates to be stabilized for structural analysis, a sample illumination protocol for cryotrapping had to be established, which was adapted from procedures reported for bacteriorhodopsin25,27,30 and channelrhodopsin-2.31 Samples

Table 1. Chemical Shifts and Torsion Angles (H-C14-C15-H) for PR in Ground and K-State and PRE108Q in M-State atom C10 C11 C12 C13 C14 C15 15 N SB H-C14-C15-H dihedral angle

ground state [ppm] K-state [ppm] M1-state [ppm] 130.0 ± 0.5 137.3 ± 0.5 130.0 ± 0.5 163.2 ± 0.5 120.2 ± 0.5 160.8 ± 0.5 181 ± 1 158° ± 2

127.6 ± 0.5 143.3 ± 0.5 123.1 ± 0.5 166.3 ± 0.5 117.7 ± 0.5 160.6 ± 0.5 181 ± 1 154° ± 3

127.0 ± 0.5 131.6 ± 0.5 127.0 ± 0.5 147.2 ± 0.5 124.1 ± 0.5 163.7 ± 0.5 316 ± 1 148° ± 3

M2-state [ppm] G-Ka [ppm] G-M1/M2b [ppm]

144.7 ± 0.5 159.4 ± 0.5 312 ± 1 160° ± 4

2.4 −6.0 6.9 −3.1 2.5 0.2 0 4°

3.0 5.7 3.0 16.0/18.5 −3.9 −2.9/1.4 −135/−131 10°/−2°

K-M1/M2c [ppm] 0.6 11.7 −3.9 19.1/21.6 −6.4 −3.1/1.2 −135/−131 6°/−6°

a

Chemical shift difference between ground and K-state. bChemical shift difference between ground and M1/M2-states cChemical shift difference between K- and M1/M2-states 16145

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

Figure 3. (A) DNP-enhanced 13C-DQF spectra of 13C2-retinal and (B) 13C6-retinal in the PR ground state (black, bottom) and upon 40 min in situ illumination of the sample in the DNP-MAS probe at ∼100 K (top). The new set of signals is assigned to the K-state. (C) Optical absorption (top) and difference spectra (bottom) of dark- and light-illuminated PR at 77 K under conditions close to the DNP experiment. The difference spectrum reveals the occurrence of the K-state.

state via thermal relaxation from the K-intermediate was not possible in PRE108Q. 13C-Spectra of 13C2-retinal in the M-state are compared with the ground and K-state in Figure 5B. In the M-state, a new C14 resonance occurs, and the K-state appears completely depopulated (Figure 5B). Both C14 and C15 resonances are broadened compared to the ground and K-states due to additional populations (see below). Although the M-state can now be trapped, the question arises whether the E108Q mutation introduces any structural changes, which would make it more difficult to relate the obtained data to the characteristics of the M-state of wild-type PR. Therefore, the 13C chemical shifts of the retinal chromophore, which serve as a structural fingerprint of the active site, were compared for PR and PRE108Q in both ground and K-states (Figures S3 and S4). All chemical shifts were found to be identical. The same observation was made for the 15 N chemical shift of the pSB (Figure 5A), which indicates that the E108Q mutation does not introduce any detectable structural changes at the chromophore. This observation fits well to the fact that position 108 is ∼15 Å away from the retinal binding pocket. Therefore, PRE108Q instead of wild-type PR is used in the following for investigating the M-state by DNPenhanced MAS NMR. DNP-Enhanced NMR Characterization of K- and MIntermediates. On the basis of the trapping procedures described above, we were able to assign and compare 13C chemical shifts of the retinal cofactor in the ground state with those observed in the K- and M-states. The assignment was derived from single bond double quantum correlations in 13 C−13C DQ-SQ41 and cross peaks in PDSD42 correlation experiments on 13C6-retinal in PR and PRE108Q as shown in Figure 6. As described above, all retinal resonances in the PR ground state could be assigned (Figures 6A, B). A number of new resonances are observed. They correlate with each other and taken together describe a nearly complete retinal signal set, which we assign to the K-state (Figure 6C, D). The reduced intensity of cross peaks with C13 is probably caused by the C20 methyl group relaxation under DNP conditions. The spectrum also contains signals from the residual ground state population. In the K-state, the chromophore of PR undergoes a trans-to-cis isomerization around the C13C14 bond due to light absorption. This results in a significant deshielding of C11 and C13 by −6.0 and −3.1 ppm, whereas C10, C12, and C14 experience an additional shielding of +2.4, 6.9, and 2.5 ppm, respectively (Table 1). Upon switching to the M-state, new

(K, M, N, O, and ground state) appear sequentially, but the population of the M-state is rather low (Figure 4C). A potential solution to this problem is offered by preventing the decay of the M-state by replacing the primary proton donor E108 with the nontitratable amino acid glutamine. The overall photodynamics of PRE108Q is elongated, and significant M-state accumulation is observed (Figure 4B). These time-resolved data are fitted with a global lifetime analysis (GLA) routine. Transient absorption changes at selected wavelengths together with their respective fits are shown for PR and PRE108Q in Figure 4C and D, respectively. According to this data analysis, the PR photocycle is best described by six lifetimes: 5 and 89 μs and 0.6, 3.4, 24.5, and 100.2 ms. The decay-associated spectra (Figure 4E) of the two fastest lifetimes describe the formation of the M-state and simultaneously the decay of the K-state. Subsequently, the M-state depopulates most likely with the two intermediate time constants of 0.6 and 24.5 ms resulting in the formation of the late intermediates. Finally, the photocycle completes with the depopulation of the N- and O-states and recovery of the ground state described by the 24.5 and 100.2 ms time constants. In contrast to wild-type PR, the GLA analysis of PRE108Q transient data results in four time constants. The respective decay associated spectra (Figure 4F) reveal a biexponential Mstate formation (70 μs and 1.2 ms) and decay (210.3 and 2261.1 ms). However, in contrast to the photodynamics of PR, the photocycle seems to end up without a detectable accumulation of the late intermediates. For this observation to be verified, an additional time-resolved absorption measurement with a temporal resolution of 1 s has been conducted (Figure S2). Even 40 s after photoexcitation, there is no indication for the formation of the late intermediates. The high population and extended lifetime of the M-state in the PRE108Q mutant should enable efficient thermal trapping of this photointermediate, which was attempted here by illumination at room temperature followed by quick freezing and transfer into the DNP probe (Figure 1B). Evidence for successful generation of the M-state in PRE108Q is provided by the detection of a 15N signal of the deprotonated Schiff base nitrogen at around 316 ppm (Figure 5A). The 15N resonance at 181 ppm indicates a remaining ground state population that is probably caused by a combination of factors such as back relaxation during the transfer step into the DNP probe but also limited quantum yield and nonideal light coupling as mentioned above. Interestingly, an accumulation of the M16146

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

Figure 4. Photocycle of wild-type PR in comparison to the PRE108Q mutant. 2D plots of transient absorption in the UV−vis spectral region of PR (A) and PRE108Q (B). Red depicts positive, blue negative, and green no absorption change. Transient absorption at 400, 500, and 580 nm of PR (C) and PRE108Q (D). The solid lines depict the fits of the data (dots). Decay-associated spectra resulting from the global lifetime analysis for PR (E) and PRE108Q (F).

Schiff base resonance. Indeed, the data in Figure 8 show that both 13C resonances of C15 correlate with two slightly different 15 N signals of the deprotonated Schiff base at 312 and 316 ppm, which overlapped in the 15N 1D spectrum (Figure 5A). Furthermore, correlations between the protonated Schiff base nitrogen and C13 and the deprotonated Schiff base and C14 are detected. No correlation between C13 and the deprotonated Schiff base could be observed because the longer distances result in a much smaller dipole coupling and therefore in a very small TEDOR signal. The observation of two M-states could be a reflection of a reorientation of the Schiff base during de- and reprotonation. Such an event could cause an additional conformational change, e.g., an out-of-plane twist, at the end of the polyene chain. To test this hypothesis, we took advantage of the 13C2-retinal (Figure 1C) and determined the H-C-C-H torsional angle

signals appear. They comprise two different sets of retinal signals with a difference observed for C13 and C15 indicating the existence of two M-states M1 and M2 (Figures 6E, F; Table 1). The largest shielding effects are observed for C11 and C13, which change by 11.7 ppm and 19.1/21.6 ppm compared to the K-state. Resonances C12 and C14 are deshielded by −3.9 and −6.4 ppm. For C15, one shielded (1.2 ppm) and one deshielded (−3.1 ppm) population is detected. The observed 13 C chemical shift changes are illustrated in Figure 7. The existence of two M-states should also be reflected in the 15 N resonance of the Schiff base nitrogen, but only one broad peak could be observed under our experimental conditions in the 1D 15N-spectrum shown in Figure 5A. Therefore, 15N−13C TEDOR dipolar through-space correlation spectra were recorded to verify that the populations identified in the 13C spectra of the retinal cofactor correlate with a deprotonated 15N 16147

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

of 154° ± 3° (Figure 9B). For the M-state, HCCH evolution curves were obtained for C15_1 and C15_2 by separating the ground state contributions by spectral deconvolution (Figure S8). Data analysis for C15_1 (163.7 ppm) reveals an angle of 148° ± 3° (Figure 9C) and for C15_2 (159.4 ppm) a value of 160° ± 4° (Figure 9D). Torsion angle data are summarized in Table 1. The HCCH torsion angle in the ground state is identical in PR and PRE108Q (Figure S5), which offers an additional verification that this mutant is structurally very similar to the wild type. Furthermore, deconvolution of the residual ground state population in K- as well as in M-state spectra results in the ground state torsion angle, verifing the robustness of the experiment and the deconvolution procedure (Figure S6).

Figure 5. (A) DNP-enhanced 15N spectra of PR and PRE108Q in ground, K-, and M-states. In the K-state, the chemical shift of the pSB at 181 ppm does not change compared to the ground state. In the Mstate, a deprotonated SB population at around 316 ppm is detected, which verifies the accumulation of the M-intermediate (* denotes a spinning sideband). (B) Comparison of DNP-enhanced 13C spectra of 13 C2-retinal in PR in the ground and K-states with 13C2-retinal in PRE108Q in the M-state. An additional set of resonances for C14 and C15 occurs. The K-state becomes completely depopulated as seen for the C14 signals (the resonance above 162 ppm is from additional isotope labeling within the protein).

■

DISCUSSION K-State. In the K-state, no 15N chemical shift change for the pSB nitrogen is observed (Figure 5A). This is in contrast to data reported for other microbial rhodopsins: Upfield shifts of 8 ppm for BR25 and 15.5 ppm for ChR231 have been reported. These shielding effects could be explained by an altered interaction between Schiff base and counterion induced by the retinal isomerization: A wealth of structural data on BR cumulates in a picture in which the highly ordered keystone water molecule 402 connects SB, Asp85, and Asp212 in the ground state. Light-induced retinal isomerization disrupts the order of this water molecule in the K-state.44 Subsequently, the H-bond between the SB and water 402 is destroyed, which would result in a 15N upfield shift. However, for PR, no corresponding water molecule has been detected so far, which could explain why the isomerization event does not affect the 15 N Schiff base chemical shift.

around the C14−C15 bond in PR and PRE108Q by double quantum spectroscopy.43 This angle reflects the orientation of the retinal polyene chain plane with respect to the Schiff base plane. In the ground state, a significant out-of-plane twist of 158° ± 2° is observed (Figure 9A), which agrees with data previously reported by us.32 For analyzing K-state data, the C14 resonance (Figure 5B) had to be deconvoluted into its ground and K-state contributions. The resulting HCCH evolution curve could be described well by a slightly altered torsion angle

Figure 6. DNP-enhanced DQSQ (A, C, E) and PDSD (B, D, F) spectra of 13C6-retinal within PR in the ground state (A, B), in the K-state (C, D), and in the M-state (E, F). The M-state was created by thermal trapping of PRE108Q. For ground-state PR, single bond double quantum correlations (ω1 + ω2) and cross peaks are highlighted by (×) in (A) and (B). In the K-state, single bond double quantum correlations (ω1 + ω2) and cross peaks are labeled with (•) in (C) and (D). Positions where residual ground state signal intensities are expected are indicated with (×). Correlations that could not be detected due to low signal intensities are labeled with (○). The M-state spectrum contains residual ground state signals (×) and two sets of single bond double quantum correlations (ω1 + ω2) and cross peaks marked with (•) and (red dots). Arrows in (F) indicate two-bond correlations between C13_1/C15_1 and C13_2/C15_2 for labeling both M-state spin systems unambiguously. Chemical shifts are summarized in Table 1. 16148

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

Figure 9. HCCH dephasing curves for the C14−C15 spin system in PR recorded over two rotor periods reporting on the HCCH dihedral angle: (A) ground state (158° ± 2°) and (B) K-state (154° ± 2°). The dephasing curve was obtained by deconvolution of the C14 resonance (Figure 5B): (C) M1-state (148° ± 4°) and (D) M2-state (160° ± 4°). The C15 resonance was deconvoluted into both M-state contributions (C151, C152) and separated from the remaining ground state population (see Figure S6 for deconvolution details).

In contrast to the Schiff base nitrogen, significant chemical shift changes are observed along the chain from C10 to C14 (Figure 7A). The largest effects are observed for C11−C13, and most remarkably, the chemical shift changes in an alternating fashion between shielding (C10, C12, C14) and deshielding (C11, C13). To separate pure isomerization effects from additional protein−retinal interactions, we overlaid known chemical shift changes from all-trans/13-cis model compounds with our experimental findings in Figure 7A (black bars). This comparison reveals that the chemical shift changes of C12 and C14 are almost identical to those in model compounds.45−47 The large shielding of C12 is a result of the reduced distance between C12 and C15 in 13-cis compared to all-trans retinal,45−47 which was also verified under our experimental conditions (Figure S7). However, the values for C10, C11, and C13 deviate significantly. These changes and the alternating shielding/deshielding pattern could be caused by an altered torsion of the polyene chain resulting in different chemical environments on both sides of the chain and/or modified protein−chromophore interactions in the K-state. The ground state crystal structure of the closely related blue PR14 shows that the retinal is sandwiched in the binding pocket between the two tryptophane residues Trp98 (helix C) and Trp197 (helix F). Furthermore, Tyr200 (helix F) is in close proximity. Carbons C11 and C13 point toward Trp197, and C10, C12, and C14 are oriented toward Trp98 (Figure S7). These residues are highly conserved and also found in BR. The consensus model for the latter involves light-induced movements of helix C toward G and a retinal interaction with Trp182 (Trp197 in green PR) pushing helix F outward.44 Time-resolved WAXS data indicate at least similar conformational dynamics also in PR.48 One could therefore speculate that the deshielding of C11 and C13 results from an increased distance and shielding of C10 by a closer proximity to Trp197 or Tyr200 in helix F (Figure S7). Moreover, additional shielding effects arising from the delocalized π-system in the retinal polyene chain and a rearrangement of bound water

Figure 7. Light-induced chemical shift changes observed for retinal carbon atoms C10−C15 in PR plotted as the difference between ground and K-state (A) and between K- and M1/M2-states (B). The black bars illustrate the effect of the all-trans-13-cis isomerization as reported for retinal model compounds45−47 (A) and the effect of deprotonation of Schiff base analogues (N-butyretinyllideneimine45) (B).

Figure 8. DNP-enhanced 15N−13C TEDOR spectrum of 13C6-retinal in 15N-Lys-PRE108Q in the M-state (tmix = 6.8 ms). Here, magnetization is transferred via TEDOR from the 15N of the pSB/SB to the retinal chain. Cross-peaks with C15 and C14 are detected in ground and Mstate populations. In the M-state, two different correlations between the SB nitrogen and C15_1 and C15_2 are observed. A correlation with C13 is only detected in the ground state. 16149

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

in a similar angle as in the ground state. A similar behavior has been observed for BR.27 The large chemical shift changes of C11−C14 during the Kto M-state transition reflect the deprotonation of the Schiff base. To which extent the deprotonation event alone dominates the experimental data can be estimated from model compound studies by comparing 13C chemical shifts of deprotonated alltrans N-butyretinylideneimine Schiff base with all-trans Nbutyretinylideneimine Schiff bases in complex with different counterions (black bars in Figure 7B).45 Our comparison reveals that the large shifts of +11.7 ppm for C11, −3.9 ppm for C12, and +19.1/21.6 ppm for C13 with respect to the K-state are fully within the range observed for these model compounds and can therefore be explained on the basis of electronic effects caused by the Schiff base deprotonation alone. Therefore, additional protein−retinal interactions can be excluded for this middle part of the polyene chain. On the contrary, the observed changes for C14 and C15 clearly deviate from those in model compounds. One reason could be found in counterion differences because, e.g., chloride and bromide are used in the model compounds in contrast to the carboxyl groups of the residues forming the counterion complex in PR. For C14 and C15_1, deshielding has been observed compared to the K-state, which is most likely caused by weaker interaction with the counterion complex due to protonation of Asp97. An additional factor could be an increased distance to Asp227, as has been reported for BR.26 These factors could be less pronounced in M2 due to the reduced chain twist resulting in a small additional shielding of C15_2. Our data for the transition from ground to K-state clearly illustrate that the changes in the retinal conformation as sampled by 13C chemical shifts and dihedral angles not only depends on the all-trans/13-cis isomerization but is strongly coupled to the protein environment. In contrast, the observed changes associated with the K to M transition are primarily controlled by the deprotonation reaction alone. Interestingly, the comparison of our experimental results with data from model compounds also suggests that the chemical shift changes induced by isomerization or deprotonation are additive. For elucidating the structural basis of the observed chemical shift changes in more detail, further studies such as quantum chemical calculations will be needed.

molecules might contribute to the observed chemical shift changes as well. Furthermore, the isomerization slightly increases the out-of-plane twist of the retinal as observed by the change of the H-C14-C15-H dihedral angle (Figure 9A, B). This reflects orientation changes between polyene chain and Schiff base and might be the first step toward the Schiff base orientation needed for the deprotonation step in the M-state. In summary, the retinal in the PR K-state shows large chemical shift changes in the middle of the polyene chain, which are different from pure isomerization effects, whereas the Schiff base linkers remain unaffected, a situation different from BR.25 M-State. We were able to trap the M-state by enhancing its low population by introducing the E108Q mutation. Our control experiments show that this mutation does not cause any structural changes within the retinal binding pocket and therefore offers a valid approach for cryotrapping the M-state. It is noteworthy that the much higher populated M-state of BR allows a relatively straightforward trapping by choosing an appropriate illumination protocol without the need of additional mutations.25,29,49 In contrast, an M-state in ChR2 could not be trapped.31 Interestingly, low populated M-states, i.e., deprotonated SB populations, are also observed in photocycles of other microbial proton pumps such as xanthorhodopsin (XR)50 and rhodopsin from Gloeobacter violaceus (GR).51 On the contrary, pumps found in a eukaryotic organism like Leptoshaeria or Neurospora rhodopsin from fungi show higher M-state populations.52,53 Any attempt to explain such differences would be highly speculative at this point, but a correlation between the M-state population and proton pumping efficiency might be assumed. However, data to verify such a hypothesis are difficult to obtain. Although BR can be considered as a “robust” proton pump, the situation is less clear for PR under physiological conditions. PR most significantly enhances growth of the host cells under starving conditions but does not increase growth of the cells in general.54−56 It has also been suggested that PR is not optimized for a fast photocycle due to the formation of the His75-Asp97 cluster, which might be an indication for alternative/additional functions as a sensor or regulator.5,33,57 The successful trapping of the PR M-state was verified by the detection of a deprotonated Schiff base 15N resonance (Figure 5A), which only appears in this part of the photocycle. The Kstate becomes completely depopulated (Figure 5B), and resonances C11−15 change significantly from the K- to Mstate (Figure 7). The observation of two C13 and C15 populations (Figure 6) and two signals for the deprotonated SB (Figure 8) provide evidence for the existence of two substates M1 (deshielding of C15) and M2 (shielding of C15), which agree well with optical data presented here (Figure 4) and elsewhere.58 The occurrence of two substates is also in accordance with similar data reported for BR.59 They could be interpreted as a reflection of a two-stage reorientation of the deprotonated SB during the photocycle from the extracellular site directly after the deprotonation to the intracellular site just before reprotonation. This is supported by the two torsion angles measured for the M-state (Figure 8): M1 could correspond to the extracellular (148°) and M2 to the intracellular orientation (160°). In M1, the SB and the polyene chain plane seem highly twisted with respect to each other. Reorientation of the SB to the intracellular site for reprotonation (M2) decreases the out-of-plane twist, resulting

■

CONCLUSIONS Here, we have presented the first structural data of the retinal chromophore in proteorhodopsin in the K- and M-states based on DNP-enhanced solid-state NMR spectroscopy. Our observations show that chemical shift changes are strongest in the middle of the retinal polyene chain and less pronounced toward the Schiff base linkage, which demonstrate a distorted retinal structure and significant protein−chromophore interactions during the ground to K-state transition. In the K-state, an alternating shielding/deshielding pattern propagating away from the Schiff base is observed, indicating that especially the middle part of the chain responds to the isomerization and experiences an altered interaction within the retinal binding pocket. The M-state shows very large chemical shift differences and splits into two populations reflecting the reorientation of the Schiff base between de- and reprotonation events, which explains the results of time-resolved optical spectroscopy. With the illumination introduced here and trapping protocol in combination with DNP-enhanced solid-state NMR, it will be possible to fully resolve the retinal binding pocket during the 16150

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

to DSS using 13C signal of adamantane at 40.49 ppm at room temperature. For all experiments, 100 kHz decoupling using SPINAL6461 was applied during acquisition. The initial 13C CP step for the proton-driven spin diffusion experiment42 (PDSD) was achieved using a contact time of 0.8 ms. A mixing time of 20 ms was used, and a recycle delay of 3 s was applied. In the indirect dimension, an acquisition time of 4.8 ms, 292 increments, and 192 scans per increment were used. 13C-double quantum filter (DQF) experiments were obtained using the POST-C762 sequence for double quantum excitation and reconversion at 0.5 ms. 2D double quantum single quantum (DQSQ) spectra41 were recorded with 254 increments in the indirect dimension with 384 scans and 17.86 μs dwell time. For 15 N−13C correlation, a TEDOR sequence with z-filter and mixing time τmix = 6.8 ms was applied.63 The 15N offset was set between the signal of the protonated and deprotonated Schiff base at 10000 Hz. Spectra were recorded with 2048 scans in direct and 64 increments with 100 μs each in indirect dimensions. Dihedral Angle Determination. For HCCH torsional angle measurements, a double quantum heteronuclear local field experiment64 was applied (HCCH experiment). In the experiment, POSTC7 during two rotor periods was used for double quantum excitation and reconversion; PMLG-965 at 106.2 kHz was used for homonuclear 1 H decoupling, and 1H CW irradiation at 106.2 kHz was used during the constant time periods. Each data point was recorded with 4096 scans. The peaks corresponding to the different photointermediates were deconvoluted with TOPSPIN 2.1 (see Figure S6). The measured peak intensities are fitted with torsion angle data, which have been simulated with SIMPSON66 as described previously.31 Time-Resolved Absorption Spectroscopy. The flash photolysis experiments have been performed with an optical parametric oscillator combined with a Nd:YAG laser (Spitlight 600, Innolas Laser GmbH) for the generation of nanosecond pump pulses. The probe pulses were generated by a Xenon flash lamp delivering microsecond white light pulses. For detection, an intensified charge-coupled device camera (ICCD camera) was deployed (PI-MAX 3, Princeton Instruments). The transient data are analyzed with a global lifetime analysis (GLA) method conducted with the program “OPTIMUS”.67 In this routine, data are analyzed based on a fixed set of exponential terms resulting in decay-associated spectra for each lifetime.68

photocycle of PR and other microbial rhodopsins. Such data are required to resolve how apparently similar architectures can encode such a diversity of functions ranging from ion pumping to gating and sensing.

■

MATERIAL AND METHODS

13

C-Labeled Retinal. 14,15- 13 C 2 -E-Retinal ( 13 C 2 -retinal), 10,11,12,13,14,15-13C6-E-retinal (13C6-retinal), and 12,1513C2-E-retinal were synthesized as described elsewhere.60 The 13C6-retinal also contained 13C-methyl groups at positions C16−C18. Because of their relaxation properties, these resonances could not be observed under DNP conditions, and they were therefore omitted. Protein Expression, Purification, and Reconstitution. PR (eBAC31A08) as well as PRE108Q were expressed in C43 E. coli cells in a defined medium containing all amino acids: 0.5 g/L of alanin, 0.4 g/ L of arginine, 0.4 g/L of aspartate, 0.05 g/l of cysteine, 0.4 g/L of glutamine, 0.65 g/L of glutamate, 0.55 g/L of glycine, 0.1 g/L of histidine, 0.23 g/L of isoleucine, 0.23 g/L of leucine, 0.2 g/L of 15Nlysine, 0.25 g/L of methionine, 0.13 g/L of phenylalanine, 0.1 g/L of proline, 2.1 g/L of serine, 0.23 g/L of threonine, 0.17 g/L of tyrosine, 0.23 g/L of valine, 0.5 g/L of adenine, 0.65 g/L of guanosine, 0.2 g/L of thymine, 0.5 g/L of uracil, 0.2 g/L of cytosine, 1.5 g/L of sodium acetate, 1.5 g/L of succinic acid, 0.5 g/L of NH4Cl, 0.85 g/L of NaOH, 10.5 g/L of K2HPO4, 20 g/L of glucose, 4 mM MgSO4, 0.05 mM FeCl3, 10 mL/L trace metal solution (20 mg of CaCl2, 20 mg of ZnSO4, 20 mg of MnSO4, 500 mg of tryptophane, 500 mg of thiamine, 500 mg of niacin, 10 mg of biotin), and 0.1 mg/L of kanamycin. In this way, specifically lysine-labeled protein could be produced. Cells were grown in 0.5 L cultures at 37 °C until OD600 = 0.6−0.8 was reached. Protein expression was induced by 1 mM IPTG. Cells were grown at 37 °C for another 3.5 h and subsequently harvested. Cell disruption and membrane preparation took place as described before.33 13C2- or 13 C6-retinal was added during solubilization in 1.5% n-dodecyl β-Dmaltoside (DDM) at 4 °C overnight. Protein was purified by Ni-NTA and reconstituted in a 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC):1,2-dimyristoyl-sn-glycero-3-phosphate (DMPA) (9:1 (w/ w)) mixture with a lipid-to-protein ratio (LPR) of 2 (w/w) as described previously.33 The proteoliposomes were washed several times with buffer (50 mM Tris, 5 mM MgCl2, pH 9) to adjust the pH and pelleted in by ultracentrifugation afterward. For signal enhancement with DNP to be achieved, the samples were incubated with 20 mM AMUPOL40 (60% D2O, 30% d8-glycerol, 10% H2O) for 14 h at 4 °C. The excess polarizing agent was removed the next day. Approximately 1.5 mg of protein was packed in a 3.2 mm sapphire rotor and spun in a rotor testing device for a few minutes at room temperature to spread the sample equally on the inner surface of the MAS rotor. Illumination and Intermediate Trapping. For illumination, high-power LEDs from Mightex were used. The K-state was trapped by direct illumination of PR samples within the DNP probe at 100 K for 40 min. The light was transferred into the probe via a glass fiber. After illumination, the sample was directly measured without further illumination. Sample transparency was found to decrease slightly at lower temperature. Therefore, the use of slightly blue-shifted LED at higher power (470 nm, 3.3 W), which maximizes light scattering through the sample, was found to yield a higher trapping efficiency compared to that of a 525 nm LED (0.13 W) (for further details, see Figure S9). The M-state was trapped by illuminating PRE108Q samples outside of the DNP probe for 1.5 min using a 525 nm LED (0.13 W) at room temperature. Subsequently, samples were flash-frozen in liquid nitrogen within the precooled rotor catcher and transferred into the DNP probe precooled at 100 K. DNP-Enhanced Solid-State NMR Experiments. All experiments were performed on a Bruker 400 DNP system consisting of a 400 MHz WB Advance II NMR spectrometer, a 263 GHz Gyrotron as microwave source, and 3.2 mm HCN Cryo-MAS probe. For 1D and 2D 13C spectra, a MAS frequency of 8 kHz was used, and for 1D 15N spectra, 10 kHz was used. During DNP experiments, the temperature was kept at around 100 K. 13C and 15N referencing was done indirectly

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b05061. Thermal relaxation data (Figure S1), transient absorption of PRE108Q (Figure S2), supporting PDSD and DQSQ spectra of PRE108Q (Figure S3), NMR data of the trapped K-state of PRE108Q (Figure S4), HCCH double quantum evolution curve for 13C2-retinal in PRE108Q (Figure S5), spectral deconvolution for HCCH analysis (Figure S6), supporting data for 12,15-13C2-E-retinal (Figure S7), supporting illustration for retinal−opsin interactions in PR (Figure S8), and optimized illumination for K-state trapping (Figure S9) (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +49-69-79829927. Fax.: +49-69-798-29929. ORCID

Alexander J. Leeder: 0000-0001-9271-6038 Josef Wachtveitl: 0000-0002-8496-8240 Clemens Glaubitz: 0000-0002-3554-6586 Notes

The authors declare no competing financial interest. 16151

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society

■

(25) Bajaj, V. S.; Mak-Jurkauskas, M. L.; Belenky, M.; Herzfeld, J.; Griffin, R. G. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 9244−9. (26) Griffiths, J. M.; Bennett, A. E.; Engelhard, M.; Siebert, F.; Raap, J.; Lugtenburg, J.; Herzfeld, J.; Griffin, R. G. Biochemistry 2000, 39, 362−371. (27) Lansing, J. C.; Hohwy, M.; Jaroniec, C. P.; Creemers, A. F. L.; Lugtenburg, J.; Herzfeld, J.; Griffin, R. G. Biochemistry 2002, 41, 431− 438. (28) Smith, S. O.; Courtin, J.; Van den Berg, E.; Winkel, C.; Lugtenburg, J.; Herzfeld, J.; Griffin, R. G. Biochemistry 1989, 28, 237− 243. (29) Hu, J. G.; Sun, B. Q.; Bizounok, M.; Hatcher, M. E.; Lansing, J. C.; Raap, J.; Verdegem, P. J.; Lugtenburg, J.; Griffin, R. G.; Herzfeld, J. Biochemistry 1998, 37, 8088−96. (30) Thompson, L. K.; McDermott, A. E.; Raap, J.; van der Wielen, C. M.; Lugtenburg, J.; Herzfeld, J.; Griffin, R. G. Biochemistry 1992, 31, 7931−8. (31) Becker-Baldus, J.; Bamann, C.; Saxena, K.; Gustmann, H.; Brown, L. J.; Brown, R. C.; Reiter, C.; Bamberg, E.; Wachtveitl, J.; Schwalbe, H.; Glaubitz, C. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 9896−901. (32) Mao, J.; Do, N. N.; Scholz, F.; Reggie, L.; Mehler, M.; Lakatos, A.; Ong, Y. S.; Ullrich, S. J.; Brown, L. J.; Brown, R. C.; Becker-Baldus, J.; Wachtveitl, J.; Glaubitz, C. J. Am. Chem. Soc. 2014, 136, 17578−90. (33) Mehler, M.; Scholz, F.; Ullrich, S. J.; Mao, J.; Braun, M.; Brown, L. J.; Brown, R. C.; Fiedler, S. A.; Becker-Baldus, J.; Wachtveitl, J.; Glaubitz, C. Biophys. J. 2013, 105, 385−97. (34) Etzkorn, M.; Seidel, K.; Li, L.; Martell, S.; Geyer, M.; Engelhard, M.; Baldus, M. Structure 2010, 18, 293−300. (35) Concistre, M.; Gansmuller, A.; McLean, N.; Johannessen, O. G.; Marin Montesinos, I.; Bovee-Geurts, P. H.; Verdegem, P.; Lugtenburg, J.; Brown, R. C.; DeGrip, W. J.; Levitt, M. H. J. Am. Chem. Soc. 2008, 130, 10490−1. (36) Kimata, N.; Pope, A.; Eilers, M.; Opefi, C. A.; Ziliox, M.; Hirshfeld, A.; Zaitseva, E.; Vogel, R.; Sheves, M.; Reeves, P. J.; Smith, S. O. Nat. Commun. 2016, 7, 12683. (37) Yomoda, H.; Makino, Y.; Tomonaga, Y.; Hidaka, T.; Kawamura, I.; Okitsu, T.; Wada, A.; Sudo, Y.; Naito, A. Angew. Chem., Int. Ed. 2014, 53, 6960−4. (38) Rosay, M.; Lansing, J. C.; Haddad, K. C.; Bachovchin, W. W.; Herzfeld, J.; Temkin, R. J.; Griffin, R. G. J. Am. Chem. Soc. 2003, 125, 13626−7. (39) Ni, Q. Z.; Daviso, E.; Can, T. V.; Markhasin, E.; Jawla, S. K.; Swager, T. M.; Temkin, R. J.; Herzfeld, J.; Griffin, R. G. Acc. Chem. Res. 2013, 46, 1933−41. (40) Sauvee, C.; Rosay, M.; Casano, G.; Aussenac, F.; Weber, R. T.; Ouari, O.; Tordo, P. Angew. Chem., Int. Ed. 2013, 52, 10858−61. (41) Hong, M. J. Magn. Reson. 1999, 136, 86−91. (42) Szeverenyi, N. M.; Sullivan, M. J.; Maciel, G. E. J. Magn. Reson. 1982, 47, 462−475. (43) Feng, X.; Lee, Y. K.; Sandstrom, D.; Eden, M.; Maisel, H.; Sebald, A.; Levitt, M. H. Chem. Phys. Lett. 1996, 257, 314−320. (44) Wickstrand, C.; Dods, R.; Royant, A.; Neutze, R. Biochim. Biophys. Acta, Gen. Subj. 2015, 1850, 536−53. (45) Harbison, G. S.; Mulder, P. P. J.; Pardoen, H.; Lugtenburg, J.; Herzfeld, J.; Griffin, R. G. J. Am. Chem. Soc. 1985, 107, 4809−4816. (46) Englert, G. Helv. Chim. Acta 1975, 58, 2367−90. (47) Inoue, Y.; Tokito, Y.; Tomonoh, S.; Chujo, R. Bull. Chem. Soc. Jpn. 1979, 52, 265−266. (48) Andersson, M.; Malmerberg, E.; Westenhoff, S.; Katona, G.; Cammarata, M.; Wohri, A. B.; Johansson, L. C.; Ewald, F.; Eklund, M.; Wulff, M.; Davidsson, J.; Neutze, R. Structure 2009, 17, 1265−75. (49) Mak-Jurkauskas, M. L.; Bajaj, V. S.; Hornstein, M. K.; Belenky, M.; Griffin, R. G.; Herzfeld, J. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 883−8. (50) Balashov, S. P.; Imasheva, E. S.; Boichenko, V. A.; Anton, J.; Wang, J. M.; Lanyi, J. K. Science 2005, 309, 2061−4. (51) Miranda, M. R.; Choi, A. R.; Shi, L.; Bezerra, A. G., Jr.; Jung, K. H.; Brown, L. S. Biophys. J. 2009, 96, 1471−81.

ACKNOWLEDGMENTS The work was funded by the Cluster of Excellence Macromolecular Complexes Frankfurt (DFG EXC 115) and DFG/ SFB 807 “Transport and communication across membranes”. The DNP experiments were enabled through an equipment grant to C.G. provided by DFG (GL 307/4-1). A.J.L., L.J.B., and R.C.D.B. acknowledge the EPSRC for funding (EP/ K039466/1 and EP/L505067/1).

■

REFERENCES

(1) Beja, O.; Aravind, L.; Koonin, E. V.; Suzuki, M. T.; Hadd, A.; Nguyen, L. P.; Jovanovich, S. B.; Gates, C. M.; Feldman, R. A.; Spudich, J. L.; Spudich, E. N.; DeLong, E. F. Science 2000, 289, 1902− 6. (2) Foster, K. W.; Saranak, J.; Patel, N.; Zarilli, G.; Okabe, M.; Kline, T.; Nakanishi, K. Nature 1984, 311, 756−759. (3) Bieszke, J. A.; Braun, E. L.; Bean, L. E.; Kang, S.; Natvig, D. O.; Borkovich, K. A. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 8034−9. (4) Finkel, O. M.; Beja, O.; Belkin, S. ISME J. 2013, 7, 448−51. (5) Man, D.; Wang, W.; Sabehi, G.; Aravind, L.; Post, A. F.; Massana, R.; Spudich, E. N.; Spudich, J. L.; Beja, O. EMBO J. 2003, 22, 1725− 31. (6) Beja, O.; Spudich, E. N.; Spudich, J. L.; Leclerc, M.; DeLong, E. F. Nature 2001, 411, 786−9. (7) de la Torre, J. R.; Christianson, L. M.; Beja, O.; Suzuki, M. T.; Karl, D. M.; Heidelberg, J.; DeLong, E. F. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 12830−5. (8) DeLong, E. F. Nature 2009, 459, 200−6. (9) Gomez-Consarnau, L.; Akram, N.; Lindell, K.; Pedersen, A.; Neutze, R.; Milton, D. L.; Gonzalez, J. M.; Pinhassi, J. PLoS Biol. 2010, 8, e1000358. (10) Friedrich, T.; Geibel, S.; Kalmbach, R.; Chizhov, I.; Ataka, K.; Heberle, J.; Engelhard, M.; Bamberg, E. J. Mol. Biol. 2002, 321, 821− 38. (11) Dioumaev, A. K.; Brown, L. S.; Shih, J.; Spudich, E. N.; Spudich, J. L.; Lanyi, J. K. Biochemistry 2002, 41, 5348−5358. (12) Inoue, K.; Kato, Y.; Kandori, H. Trends Microbiol. 2015, 23, 91− 8. (13) Wang, W. W.; Sineshchekov, O. A.; Spudich, E. N.; Spudich, J. L. J. Biol. Chem. 2003, 278, 33985−91. (14) Ran, T.; Ozorowski, G.; Gao, Y.; Sineshchekov, O. A.; Wang, W.; Spudich, J. L.; Luecke, H. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2013, 69, 1965−80. (15) Maciejko, J.; Mehler, M.; Kaur, J.; Lieblein, T.; Morgner, N.; Ouari, O.; Tordo, P.; Becker-Baldus, J.; Glaubitz, C. J. Am. Chem. Soc. 2015, 137, 9032−43. (16) Mors, K.; Roos, C.; Scholz, F.; Wachtveitl, J.; Dotsch, V.; Bernhard, F.; Glaubitz, C. Biochim. Biophys. Acta, Biomembr. 2013, 1828, 1222−9. (17) Bamann, C.; Bamberg, E.; Wachtveitl, J.; Glaubitz, C. Biochim. Biophys. Acta, Bioenerg. 2014, 1837, 614−25. (18) Shi, L. C.; Ahmed, M. A. M.; Zhang, W. R.; Whited, G.; Brown, L. S.; Ladizhansky, V. J. Mol. Biol. 2009, 386, 1078−1093. (19) Eckert, C. E.; Kaur, J.; Glaubitz, C.; Wachtveitl, J. J. Phys. Chem. Lett. 2017, 8, 512−517. (20) Reckel, S.; Gottstein, D.; Stehle, J.; Lohr, F.; Verhoefen, M. K.; Takeda, M.; Silvers, R.; Kainosho, M.; Glaubitz, C.; Wachtveitl, J.; Bernhard, F.; Schwalbe, H.; Guntert, P.; Dotsch, V. Angew. Chem., Int. Ed. 2011, 50, 11942−6. (21) Yang, J.; Aslimovska, L.; Glaubitz, C. J. Am. Chem. Soc. 2011, 133, 4874−81. (22) Lanyi, J. K. Mol. Membr. Biol. 2004, 21, 143−50. (23) Bergo, V. B.; Sineshchekov, O. A.; Kralj, J. M.; Partha, R.; Spudich, E. N.; Rothschild, K. J.; Spudich, J. L. J. Biol. Chem. 2009, 284, 2836−43. (24) Hempelmann, F.; Holper, S.; Verhoefen, M. K.; Woerner, A. C.; Kohler, T.; Fiedler, S. A.; Pfleger, N.; Wachtveitl, J.; Glaubitz, C. J. Am. Chem. Soc. 2011, 133, 4645−54. 16152

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153

Article

Journal of the American Chemical Society (52) Brown, L. S.; Dioumaev, A. K.; Lanyi, J. K.; Spudich, E. N.; Spudich, J. L. J. Biol. Chem. 2001, 276, 32495−505. (53) Fan, Y.; Shi, L.; Brown, L. S. FEBS Lett. 2007, 581, 2557−61. (54) Fuhrman, J. A.; Schwalbach, M. S.; Stingl, U. Nat. Rev. Microbiol. 2008, 488−494. (55) Gomez-Consarnau, L.; Gonzalez, J. M.; Riedel, T.; Jaenicke, S.; Wagner-Dobler, I.; Sanudo-Wilhelmy, S. A.; Fuhrman, J. A. ISME J. 2016, 10, 1102−12. (56) Palovaara, J.; Akram, N.; Baltar, F.; Bunse, C.; Forsberg, J.; Pedros-Alio, C.; Gonzalez, J. M.; Pinhassi, J. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, E3650−8. (57) Sabehi, G.; Beja, O.; Suzuki, M. T.; Preston, C. M.; DeLong, E. F. Environ. Microbiol. 2004, 6, 903−10. (58) Varo, G.; Brown, L. S.; Lakatos, M.; Lanyi, J. K. Biophys. J. 2003, 84, 1202−7. (59) Herzfeld, J.; Lansing, J. C. Annu. Rev. Biophys. Biomol. Struct. 2002, 31, 73−95. (60) Leeder, A. J.; Brown, L. J.; Becker-Baldus, J.; Mehler, M.; Glaubitz, C.; Brown, R. C. D. J. Labelled Compd. Radiopharm. 2017, accepted, DOI: 10.1002/jlcr.3576. (61) Fung, B. M.; Khitrin, A. K.; Ermolaev, K. J. Magn. Reson. 2000, 142, 97−101. (62) Hohwy, M.; Jakobsen, H. J.; Eden, M.; Levitt, M. H.; Nielsen, N. C. J. Chem. Phys. 1998, 108, 2686−2694. (63) Jaroniec, C. P.; Filip, C.; Griffin, R. G. J. Am. Chem. Soc. 2002, 124, 10728−10742. (64) Concistre, M.; Johannessen, O. G.; McLean, N.; Bovee-Geurts, P. H.; Brown, R. C.; Degrip, W. J.; Levitt, M. H. J. Biomol. NMR 2012, 53, 247−56. (65) Vinogradov, E.; Madhu, P. K.; Vega, S. Chem. Phys. Lett. 1999, 314, 443−450. (66) Bak, M.; Rasmussen, J. T.; Nielsen, N. C. J. Magn. Reson. 2000, 147, 296−330. (67) Slavov, C.; Hartmann, H.; Wachtveitl, J. Anal. Chem. 2015, 87, 2328−36. (68) van Stokkum, I. H.; Larsen, D. S.; van Grondelle, R. Biochim. Biophys. Acta, Bioenerg. 2004, 1657, 82−104.

16153

DOI: 10.1021/jacs.7b05061 J. Am. Chem. Soc. 2017, 139, 16143−16153